Introduction

¹⁹F MRI of labeled stem cells following intra-cardiac injections has been successfully implemented in rodents with conjugated (protamine-sulphate (PS), NP1) [1] or non-conjugated (NP2) poly(lactic-glycolic-acid)-coated perfluorocarbon nanoparticles (PFC-NPs, NP2), primarily based on either PS-mediated endocytosis [2] or FuGENE-encapsulation and uptake [3]. However, a direct comparison of the labeling capacity of NP1/NP2 for achieving direct cardiac ¹⁹F MRI on the same animal species is still lacking. Therefore, the aim of this study is to compare the labeling capacities of NP1/NP2 in vitro, and in the mouse in vivo, based on ¹⁹F MRI of labeled cardiac progenitor cells (CPCs).

Methods

Cell Isolation/Labeling: CPCs were isolated from adult, C57BL/6 mouse atria, and labeled using NP1 (1.5 mg/ml, U. Coimbra, Portugal) or NP2 (10 mg/ml, Atto647, Radboud U., Netherlands) for ~24 h, as described previously [3, 4]. To increase the label uptake, a successful electroporation protocol was implemented for NP1 into CPCs with a Gene Pulser Xcell electroporator (BIORAD, Hertfordshire, UK) using a rectangular pulse (amplitude 250 V, duration 5 ms), followed by NP1 incubation for ~1.5 h. In Vivo Cardiac/Post-mortem Skeletal Muscle Imaging: Two mice were injected with either NP1 (mixture of labeled/electroporated) or NP2-labeled CPCs (~3 and ~1×10⁶ CPCs) in femoral areas and imaged post-mortem. Another mouse was injected with NP2-labeled cells (~1.7-2×10⁶ CPCs) under 2% isoflurane (ISO) and 100% oxygen followed by in vivo ¹⁹F MRI. MRI/MRS: All experiments were conducted on a 9.4 T Agilent scanner (Agilent Technologies, USA) using a 40×20 mm² butterfly (post-mortem, in vivo), and a homogeneous 5 (diameter) × 8 (length) mm² solenoid (in vitro concentration comparisons and T₁/T₂ measurements). The broad frequency response of the coils allowed intermittent ¹H/¹⁹F MRI/MRS. Phantom Studies: A phantom containing NP1 media solutions in Eppendorf vials at different concentrations (1-10 mg/ml) and a trifluoroacetic acid (TFA) solution (75mM) was studied (butterfly coil) using ¹H/¹⁹F MRI/MRS (1H spoiled gradient echo (SPGR): repetition time (TR)/echo time (TE)=64.3/4/number of excitations (NEX)=6/flip angle=20°/slice thickness (ST)=5 mm/field-of-view (FOV)=40×40 mm²/bandwidth (BW)=100 kHz/19F MRI (2D-SPGR)/TR/TE=5.8-5.9/2.9ms/NEX=256/flip angle=50°/ST=5 mm/BW=6kHz; ¹⁹F MRS (pulse-acquire): TR=100ms/512 points/BW=15kHz/NEX=512), and the results were compared with NP2s, as reported previously [4]. Fluorine relaxation measurements (¹⁹F: Rd-NP solutions: NP1-labeled and electroporated-labeled CPCs), and ¹⁹F concentration measurements, were conducted using the homogeneous solenoid coil using pulse-acquire, inversion recovery (T₁), or Carr-Purcell-Meiboom-Gill sequences (T₂). In Vivo Cardiac/Post-mortem Skeletal ¹⁹F MRI: All studies used the butterfly coil. For the in vivo cardiac MRI, the mouse was anesthetized with 4% ISO and maintained and monitored for up to 95 min with 1.5–2.0% ISO in 100% oxygen. ¹H Imaging: Double-gated, two-dimensional (2D) multislice, and ungated 2D/3D cardiac ¹H images of the mouse thorax were acquired in vivo (2D: TR/TE=1.9-2.13/1-1.1 ms/flip angle=20-50°/NEX=2-6/FOV=40×40 mm²/matrix= 128×128/acquisition time=3-5 min; 3D: TR/TE=1.9-2.13/1-1.1 ms/flip angle=20-50°/NEX=2-6, FOV=40×40 mm²/ matrix= 128×128×128/acquisition time=3-5 min). ¹⁹F MRI/MRS: Ungated 19F image acquisitions matched ¹H acquisitions (2D: TR/TE=16.5/8.3 ms/flip angle= 50°/NEX=768/FOV=40×40 mm²/1 slice/ST=5 mm/matrix=32×32/BW=2–4 kHz/acquisition time=2.42 min/and 8 phase encoding steps per segment; 3D: TR/TE=16.4/8.3 ms/flip angle=20-30°/NEX=12-72/FOV=40×40 mm²/matrix= 32×32×32/BW=1-3 kHz/acquisition time=3-6 min). Similar imaging parameters /protocols were used for post-mortem skeletal muscle imaging (NEX=2048 for NP1 and 48 for NP2). Image Processing: ¹⁹F images were imported and interpolated in ImageJ (NIH, Bethesda, USA) to match the ¹H matrix size, and overlaid with ¹H MRI (opacity=30-70%). Spectra were processed in CSX (Johns Hopkins, USA) or in IDL (Harris Geospatial, USA).

Results

Fast acquisitions confirm detectability of NP1 solutions only at doses of 10 mg/ml (Fig. 1), while the concentration of the fluorine cell label is ~10-50 times smaller compared to NP2 (Fig. 2). Relaxation values are significantly shorter for NP1 (labeled and electroporated) compared to NP2 (T₁, p<0.02 and p<0.03; T₂, p<0.04), while electroporation leads to significant decreases in T₂ values, thereby prohibiting their measurements (Fig. 2). Post-mortem skeletal muscle imaging of administered CPCs confirms visibility for both label types. However, cardiac ¹⁹F MRI visualization was only possible with NP2, despite successful injection protocols for both labels (Fig. 3).

Discussion-Conclusion

We demonstrate improved cardiac ¹⁹F MRI performance for NP2 compared to NP1. The decreased relaxation times of NP1 elicit minor benefits for fast imaging acquisitions indicating that the differences in cardiac MRI visibility for the two labels are attributed to their ¹⁹F content or uptake mechanism. Increase of the fluorinated content of NP1 is envisaged to lead to comparable MRI visibility as that for NP2. Relaxation measurements for NP1 after the fluorine content is increased will explain whether the observed differences are owing to PFC content, or to PS-complexation and the cellular uptake process—compared to FuGENE-based encapsulation uptake for NP2.

Acknowledgements

The work was supported by the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie [grant agreement No. 652986], and the European Research Council Grant ERC-2013-StG-336454 (MS).